Sensitive detection of miRNA by using hybridization chain reaction coupled with positively charged gold nanoparticles

Positively charged gold nanoparticles (+)AuNPs can adsorb onto the negatively charged surface of single-stranded DNA (ssDNA) or double-stranded DNA (dsDNA). Herein, long-range dsDNA polymers could form based on the hybridization chain reaction (HCR) of two hairpin probes (H1 and H2) by using miRNA-21 as an initiator. (+)AuNPs could adsorb onto the negatively charged surface of such long-range dsDNA polymers based on the electrostatic adsorption, which directly resulted in the precipitation of (+)AuNPs and the decrease of (+)AuNPs absorption spectra. Under optimal conditions, miRNA-21 detection could be realized in the range of 20 pM-10 nM with a detection limit of 6.8 pM. In addition, (+)AuNPs used here are much more stable than commonly used negatively charged gold nanoparticles ((−)AuNPs) in mixed solution that contained salt, protein or other metal ions. Importantly, the assay could realize the detection of miRNA in human serum samples.

The key point in colorimetric-based bioanalysis is to improve the sensitivity of such proposals. Recently, numerous amplification strategies including rolling circle amplification [35][36][37] , enzymatic signal amplification [38][39][40] and strand displacement assay 41 have developed. However, these amplified strategies suffered from some weaknesses including complex process, high cost and possible false response because of the utilization of enzyme 42 . Therefore, hybridization chain reaction (HCR), as one type of enzyme-free amplification strategies, can be used as an ideal candidate, which is driven by the self-assembly of two stable species of DNA hairpins 43 and can also exhibit the advantages of simple operation, low background, low cost and PCR-like sensitivity 44 .
Due to the unique optical properties and high molar extinction coefficient of gold nanoparticles (AuNPs) 45 , they have been widely used for the preparation of colorimetric biosensors, based on detecting the color change of AuNPs from red to blue (assembly) or from blue to red (disassembly) induced by the distance change of particles [46][47][48] . AuNPs used in such sensing strategies are mainly negatively charged and the disperse states of them can be significantly affected in complex systems that contained salt, DNA, proteins or other ions, which limited the application of them in biological and complex systems assay. Recently, researchers reported several novel colorimetric-based strategies by using positively charged AuNPs ((+ )AuNPs) as signal probe based on the Scientific RepoRts | 6:32358 | DOI: 10.1038/srep32358 adsorption of them to negatively charged surface of DNA 49,50 . Such label-free methods are relatively simple and low-cost, without needing the modification of substrate with AuNPs.
Here we developed for the first time a versatile and sensitive label-free sensor for miRNA-21 detection based on the precipitation of (+ )AuNPs, and such precipitation happened automatically along with the increase of miRNA-21 concentration, which without needing an extra magnet just as the literatures [51][52][53] . Meantime, such proposed strategy avoided the lengthy conjugation process between DNA and AuNPs. In addition, experiments testified that (+ )AuNPs used here are much more stable than negatively charged AuNPs in mixed systems containing salt, DNA, protein or metal ions (See result and discussion part), making our proposed method more tolerant to the sensing environments. Moreover, the assay contributed the new application of (+ )AuNPs and also expanded the scope of miRNA based sensing. Finally, a satisfactory result obtained while detecting miRNA-21 in human serum samples.

Results and Discussion
Detection mechanism of miRNA-21. To construct such a sensor, HCR happened between two hairpin probes (H 1 and H 2 ) by using the target miRNA-21 as an initiator to form long-range dsDNA polymers. Then, numerous (+ )AuNPs could adsorb onto the surface of such long-range dsDNA polymers based on the electrostatic adsorption, and accordingly resulted in the precipitation of them because of the increase of their gravity. Meantime, more HCR products could be formed along with the increase of miRNA concentration, which directly induced the precipitation of (+ )AuNPs more easily as a result. At last, miRNA-21 could be detected facilely based on detecting the concentration of (+ )AuNPs in supernatant by using the simple UV-vis spectrum method ( Fig. 1).
Characterizations of (+)AuNPs. UV-vis spectrum of (+ )AuNPs in supernatant were performed in order to investigate the characteristics of the proposed method. Figure 2A showed the UV-vis spectrum and the corresponding photographs of (+ )AuNPs. In the absence of target miRNA-21, a high adsorption spectra of (+ )AuNPs was observed at 510 nm (a), and no precipitation was found (inset, a). However, in the presence of 1.0 and 5.0 nM of target miRNA-21 (b and c), the adsorption spectra of (+ )AuNPs at 510 nm decreased obviously, accompanying with the precipitation of them (inset, b and c). Such results directly illustrated that HCR happened between H 1 and H 2 to form long-range dsDNA polymers for the precipitation of (+ )AuNPs. The reason for such precipitation might be due to the increased specific gravity of DNA-AuNPs nanostructure. Literatures have reported that the change of AuNPs color from orange red to blue was mainly because of the increase of particle size or the massive aggregate of AuNPs 54,55 . Herein, the color of (+ )AuNPs did not change only with a precipitation. Thus, it could be supposed that the precipitation of (+ )AuNPs was mainly due to the adsorption of them with DNA strands 56 but not the production of massive aggregates.
Meantime, calf thymus DNA (ctDNA, dsDNA) was selected to further prove the adsorption properties of (+ )AuNPs on negatively charged dsDNA strands. As seen in Fig. 2B (a), (+ )AuNPs precipitation appeared upon the incubation of them with ctDNA (0.2 mg/mL), which was similar to the phenomenon of that (+ )AuNPs were incubated with long-range dsDNA polymers (b). Such results further illustrated that the precipitation of (+ ) AuNPs was based on the adsorption of them with long-range dsDNA polymers.
The total size of (+ )AuNPs increased after the incubation of them with long-range dsDNA polymers. As shown in Fig. 3A, the average hydrodynamic diameter of (+ )AuNPs detected from dynamic light scattering (DLS) was 11.2 nm (a), which was bigger than the TEM results (insert a) because of that the DLS analysis measured the hydrodynamic radius while TEM analysis provided a more precise measurement of the hard AuNP core 57 . Then, upon the incubation of (+ )AuNPs with the long-range dsDNA polymers formed between 300 nM of H 1 /H 2 and 5.0 nM of target miRNA-21, (+ )AuNPs assembly and precipitation happened and the average hydrodynamic diameter of (+ )AuNPs increased to 146.8 nm (b). Such results were in good agreement with the UV-vis absorption spectra results in Fig. 2A.  HCR is a critical factor for constructing such a sensor. Thus, gel electrophoresis was conducted to monitor the happen of HCR between H 1 and H 2 by using the target miRNA-21 as an initiator. As shown in Fig Moreover, the zeta potential analysis of (+ )AuNPs was also constructed by using DLS to investigate the charge reduction of (+ )AuNPs before and after the incubation of them with long-range dsDNA polymers. As seen in Fig. 3B, the zeta potential of (+ )AuNPs was + 39.8 mV (a) while it reduced to + 15.1 V after the incubation of it with H 1 and H 2 mixture in the absence of miRNA-21, which mainly due to the neutralization of (+ )AuNPs charge by negatively charged H 1 and H 2 DNA strands (b). Then, the zeta potential reduced to + 10.1 V (c) upon the incubation of (+ )AuNPs with long-range dsDNA polymers formed between 300 nM of H 1 /H 2 and 5.0 nM of target miRNA-21. Such reduction of (+ )AuNPs charge mainly due to the effective adsorption of them to negatively charged long-range dsDNA polymers.
Optimization of the experimental conditions. To investigate the size effect of (+ )AuNPs on the performance of the sensor, three types of (+ )AuNPs samples (4 nm, 16 nm and 32 nm) were prepared and the UV-vis absorption spectra of them were similar (Fig. 4A). Then, three types of (+ )AuNPs were incubated with H 1 and H 2 mixture (300 nM, 1:1 ratio) respectively in the presence of 5.0 nM of target miRNA-21. As shown in inset of Fig. 4B, the degree of (+ )AuNPs precipitation was decreased along with the increase of (+ )AuNPs size (a: 4 nm, b: 16 nm and c: 32 nm), followed by the increase of (+ )AuNPs absorption spectra (Fig. 4B). Such phenomena indicated that 4 nm of (+ )AuNPs in favor of the precipitation of AuNPs-DNA nanostructure. Thus, 4 nm of (+ ) AuNPs was selected as the signal probe for all of the experiments.
To prove the stability of (+ )AuNPs in mixed solution, the effect of salt (NaCl as example), protein (BSA as example), metal ions (K + , Ca 2+ and Mg 2+ as example) and DNA probes concentration was investigated. As shown in Fig. 4C, no obvious change of (+ )AuNPs absorption spectra appeared along with the increase of NaCl (a), BSA (b) and metal ions (c) concentration from 0 to 1.0 M, respectively. Such results certainly indicated that (+ )AuNPs are highly stable in mixed environments. However, if the concentration of H 1 and H 2 was higher than 300 nM, (+ )AuNPs precipitation happened because of the absorption of them with DNA strands. Thus, 300 nM of H 1 /H 2 was used in this sensor.
Moreover, the incubation time of (+ )AuNPs with H 1 and H 2 mixture (300 nM) in the absence and presence of target miRNA-21 was investigated in Fig. 4D. The adsorption spectra of (+ )AuNPs had slight change against the incubation time in the absence of miRNA-21 (a), which mainly due to the adsorption of a small number of (+ )AuNPs with H 1 and H 2 strands. However, in the presence of 5.0 nM miRNA-21, the adsorption spectra of (+ )AuNPs gradually decreased with increasing incubation time over the range from 0-25 min (b), and then reached a plateau. Results indicated that 25 min was enough for the effectively adsorption of (+ )AuNPs with negatively charged dsDNA polymers. Thus, 25 min was selected as the optimum incubation time.
Performance of the sensor. Properties of the proposed method was investigated based on detecting the change of (+ )AuNPs absorption spectra at 510 nm upon the incubation of them with H 1 and H 2 mixture (300 nM, 1:1 ratio) in the presence of different concentrations of target miRNA-21. As shown in Fig. 5A, (+ )AuNPs precipitation increased along with the increase of target miRNA-21 concentration, and followed by the decrease of (+ )AuNPs absorption spectra at 510 nm (Fig. 5B). A linear range of miRNA-21 concentration from 20 pM to 10 nM was obtained with a detection limit of 6.8 pM (containing 1.7 × 10 6 /μ L copies) recorded using the 3σ method (Fig. 5C). The linear regression equation was A = 2.317-0.016 c (c: nM, R 2 = 0.997). Moreover, the detection limit of our assay could compare with some nanoparticle-based optical methods because of the HCR signal amplification of our method 18,61,62 . Meantime, our method was simple, without needing the using of enzyme and expensive instrument, which made our sensor more suitable for bioanalysis than other methods by using enzyme (Table 1) 16,63 . Selectivity of the sensor. To investigate the sequence specificity of the assay, we designed two types of mismatch RNA strands, named as mismatch 1 and mismatch 2, respectively. One U base in mismatch 1 was replaced by C while two U bases in mismatch 2 were replaced by CG. The selectivity of the sensor for miRNA-21 detection was investigated by monitoring the UV-vis absorption spectra change (Δ A) of (+ )AuNPs after incubating them with H 1 and H 2 mixture in the presence of miRNA-21, mismatch 1 and mismatch 2, respectively. As shown in Fig. 6A, the Δ A of (+ )AuNPs induced upon the addition of 5.0 nM of miRNA-21 was 1.0, while they were 0.27 and 0.03 for mismatch 1 and mismatch 2. Such results directly indicated that the proposed sensor has high selectivity for miRNA detection.

Application of the sensor in real samples.
To evaluate the application of the sensor for miRNA-21 detection in real samples, human serum samples were collected from Xuzhou Central Hospital. Then, the recovery experiments were constructed using standard addition method by adding 50 pM, 500 pM and 5.0 nM of miRNA-21 to human serum samples. As shown in Table 2 a good recovery in the range of 97.8-101.3% was obtained, and the relative standard deviation (RSD) was below 4.63%. Meantime, the concentrations of miRNA-21 in three human cancer cell lines were detected by using our assay and qRT-PCR method. As shown in Fig. 6B, the detection results obtained by using our assay were consistent with those of the qRT-PCR method, and each MCF-7 cell contains ~382 copies of miRNA-21, which was similar to the result of the literature 64 . Such results indicated that the proposed sensing platform has a potential application for the detection of miRNA-21 in complex biological samples.
In conclusion, a versatile and label-free sensor for target miRNA-21 detection was developed based on the precipitation of (+ )AuNPs. Combining (+ )AuNPs with HCR dual signal amplification, a detection limit down to 6.8 pM was obtained for target miRNA-21, which was much more sensitive compared with other colorimetric based methods. In addition, such visual based detection method was simple and rapid, without needing the label of DNA strands and using any experiment instruments. Moreover, (+ )AuNPs, used in such strategy, has good stability in mixed solution that contained salt, protein or other metal ions. These advantages make our method more convenient for the application of biochemical and biomolecules detection.  Table 3. The 22-nt bases of H 1 at the 5′ end (italic) were used to hybridize with miRNA-21, and 24-nt bases of H 1 at the 3′ end (blue underline) were used to hybridize with 24-nt bases of H 2 at the 3′ end (red underline). Then, the 22-nt bases of H 2 at the 5′ end could hybridize with another H 1 strands, and further induced the occurrence of HCR between H 1 and H 2 . Before using, H 1 and H 2 were heated to 95 °C and stayed for 5 min, then allowed to cool to room temperature in one hour, and the anneal    product was stored in a refrigerator at 4 °C before use. Cetyltrimethyl ammonium bromide (CTAB), chloroauric acid (HAuCl 4 ) and sodium borohydride (NaBH 4 ) were purchased from Aladdin Biotech CO. Ltd. (Beijing, China). Ultra-pure water was obtained from Heal Force Smart-Nultra-pure water system and used for all of the experiments. All other chemicals were analytical grade and used without further purification. 50 mM of NaAc-HAc (pH 5.0) was used for DNA and miRNA dissolution, 50.0 mM of tris-HCl (pH 7.4, with 50 mM of NaCl) was used for miRNA-21 detection. UV-vis absorption spectra of (+ )AuNPs were carried out on an evolution 60 spectrophotometer (Thermo Fisher Corporation, USA). The size distribution of (+ )AuNPs was characterized by transmission electron microscope (TEM, JEM-2010HR, Japan). The zeta potential analysis of (+ )AuNPs was investigated using particle size analyzer (MS2000, England).

Preparation and characterization.
(+ )AuNPs were prepared according to our previous method 65 . Firstly, 15 mL of HAuCl 4 (1.0 mM) and 2 mL of CTAB (10 mM) were mixed and stirred for 15 min. Subsequently, 2 mL of NaBH 4 (100 mM) was added to the mixed solution and kept stirring for another 8 min. After the solution color changed to orange red and without changing within 15 min, the solution was filtered and stored in a refrigerator at 4 °C before use. The average size of such (+ )AuNPs was about 4 nm estimated from TEM images (Fig. 3A(insert a)). Meantime, the zeta potential of such nanoparticles monitored form DLS were positively charged ( Fig. 3B(a)).
To Visual detection of miRNA-21. H 1 and H 2 were firstly mixed and heated to 95 °C for 5 min, followed by cooling to room temperature gradually, and then they were stored at 4 °C prior to use. For miRNA-21 detection, different concentrations of target miRNA-21 samples were mixed with 10 μ L of H 1 and H 2 mixture (300 nM, 1:1 ratio) separately, followed by incubating at 37 °C for 1 h to proceed the HCR. After that, 100 μ L of (+ )AuNPs (10.6 nM) were added into above solutions respectively and incubated for another 25 min. Due to the electrostatic adsorption of long-range DNA polymers with (+ )AuNPs, (+ )AuNPs precipitation appeared. Finally, the concentrations of (+ )AuNPs in supernatant of each above mixed solutions were collected and detected by using UV-vis characterization.
Visual detection of miRNA-21 in real samples. To evaluate the application of the sensor for miRNA-21 detection in real samples, human serum samples were collected from Xuzhou Central Hospital and diluted for 10 times with PBS buffer before using. Then, 50 pM, 500 pM and 5.0 nM of miRNA-21 samples were mixed with 10 μ L of H 1 and H 2 mixture (300 nM, 1:1 ratio) separately, and then added them to human serum samples followed by incubating at 37 °C for 1 h to proceed the HCR. After that, 100 μ L of (+ )AuNPs (10.6 nM) were added into above solutions respectively and incubated for another 25 min. Finally, the concentrations of (+ )AuNPs in supernatant of each above mixed solutions were collected and detected by using UV-vis characterization, and the concentration of miRNA was found from the calibration curve.
Cell culture and total RNA extraction. Three types of human cancer cell lines including human breast cancer cell line (MCF-7), human cervical carcinoma cell line (HeLa) and human mammary epithelial cell line (MCF-10A) were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 15% fetal bovine serum (FBS), 100 U/mL penicillin and 100 g/mL streptomycin at 37 °C in a humidified atmosphere containing 5% CO 2 . The cellular extracts were prepared according to the literature method 64 .